bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.145136; this version posted June 11, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Research Article Sex-dependent effects of chronic exercise on cognitive flexibility in aging mice Running title: Cognitive flexibility in aging mice Annabel K. Short1†, Viet Bui1,2, Isabel C. Zbukvic1‡, Anthony J. Hannan1,2, Terence Y. Pang1,2, Jee Hyun Kim*1,2 1. Florey Department of Neuroscience and Mental Health, University of Melbourne, Parkville, VIC 3052 Australia 2. Mental Health theme, The Florey Institute of Neuroscience and Mental Health, Parkville, VIC 3052 Australia *Correspondence: Associate Professor Jee Hyun Kim Email: [email protected] Mental Health Theme The Florey Institute of Neuroscience and Mental Health Kenneth Myer Building 30 Royal Pde Parkville, VIC 3052 Australia Phone: +61-3-9035-6623 Footnote (present address): †Department of Pediatrics, University of California – Irvine, California 92617 ‡ Orygen, Parkville, Victoria, Australia; Centre for Youth Mental Health, The University of Melbourne, Parkville, Victoria, Australia 1 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.145136; this version posted June 11, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Abstract Cognitive impairments associated with advanced age are a growing concern in our aging society. Such impairments are associated with alterations in brain structure and function, especially in the hippocampus, which changes to experience throughout life. It is well-known that regular exercise can maintain hippocampus volume. The hippocampus is critical for cognitive flexibility involved with extinction and reinstatement of conditioned fear. Therefore, we asked whether voluntary chronic exercise in middle-aged mice can improve extinction and/or reinstatement of conditioned fear compared to standard housing. Eight-month-old male and female C57Bl/6J mice had access to a running wheel or remained in standard housing until 11 months of age. Alongside control standard-housed young adult (3-month- old) mice, they received tone-footshock pairings, which were subsequently extinguished with tone-alone presentations the next day. Half of the mice then received a reminder treatment in the form of a single footshock. Both male and female 11-month-old mice housed in standard conditions exhibited impaired reinstatement compared to young adult mice. However, for males that had access to a running wheel from 8 months of age, the reminder treatment rescued reinstatement ability. This was not observed in females. Additionally, exercise during middle age in both sexes increased expression of Bdnf mRNA in the hippocampus, specifically exon 4 mRNA. These results show that, at least for males, physical exercise is beneficial for reducing age-related decline in cognitive abilities. Despite not rescuing their impaired reinstatement, exercise also increased Bdnf gene expression in the female hippocampus, which could potentially benefit other forms of hippocampal- dependent cognition. Keywords: sex characteristics; cognition; exercise; hippocampus; Bdnf; reinstatement 2 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.145136; this version posted June 11, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1. Introduction Aging is associated with reduced executive functions, with cognitive flexibility being one of the most impaired facets of intelligence due to age (Salthouse, 1996). Cognitive flexibility is the ability to adapt to a changing environment and is often tested using reversal or set shifting tasks (Dajani & Uddin, 2015; Scott, 1962). Currently, medical interventions for cognitive impairments associated with aging are limited. Thus, identifying new ways to delay age-associated cognitive problems is vital in our aging society. People with a history of a physically active lifestyle are known to have some resilience to the effects of normal aging on cognition. In humans, the correlation between aerobic exercise and improved cognition is strong and exercise appears to be effective across the lifespan (Chaddock et al., 2010; Erickson et al., 2011; Herting & Nagel, 2012; Kleemeyer et al., 2016; Rosano et al., 2017; Stillman et al., 2018; Thomas et al., 2016). This has significant implications for older adults who face heightened risk of dementia (Erickson et al., 2011; Kleemeyer et al., 2016; Rosano et al., 2017). The effects of exercise on cognition has been linked with hippocampal function (Rubin, Watson, Duff, & Cohen, 2014), which is important for spatial tasks and cognitive flexibility (Burghardt, Park, Hen, & Fenton, 2012). Although many studies highlight changes in the hippocampus volume and connectivity as the neural correlate for exercise effects in humans (Chaddock et al., 2010; Erickson et al., 2011; Herting & Nagel, 2012; Kleemeyer et al., 2016; Rosano et al., 2017), the molecular correlates are poorly understood. Rodent models have been useful in this regard and have provided additional insights. For example, chronic exercise can alleviate the decrease of hippocampal neurogenesis and synaptic plasticity in aging rodents (Anacker & Hen, 2017; van Praag, Shubert, Zhao, & Gage, 2005). Importantly, the hippocampus is a sexually-dimorphic brain region in rodents and humans, dependent on hormonal cycle stage (reviewed in Yagi & Galea, 2019). Indeed, sex differences in hippocampus-dependent learning have been widely reported in rodents and humans, with males typically showing superior spatial learning (Jonasson, 2005; Linn & 3 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.145136; this version posted June 11, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Petersen, 1985; Voyer, Voyer, & Bryden, 1995). However, whether there are sex differences in cognitive decline with age is unclear in humans (Ferreira, Ferreira Santos-Galduróz, Ferri, & Fernandes Galduróz, 2014; Karlsson, Thorvaldsson, Skoog, Gudmundsson, & Johansson, 2015; McCarrey, An, Kitner-Triolo, Ferrucci, & Resnick, 2016; Workman, Healey, Carlotto, & Lacreuse, 2019; Zaninotto, Batty, Allerhand, & Deary, 2018) . In rodents, there are very few reports of sex differences in non-pathological cognitive decline with age (Zanos et al., 2015). Sex differences are observed following exercise in hippocampus-dependent tasks, although rodent and human findings differ. A meta-analysis in humans reported sex-differences in the level of cognitive improvements following exercise, with females having greater improvements especially after aerobic training (Barha, Davis, Falck, Nagamatsu, & Liu- Ambrose, 2017). A similar meta-analysis in rodents describes no sex differences in spatial tasks following aerobic training, but greater improvements in non-spatial cognitive tasks in males (Barha, Falck, Davis, Nagamatsu, & Liu-Ambrose, 2017). Taken together, it is clear that sex-specific effects on cognition and underlying neurobiology need further examination. The aim of the present study was to examine the impact of exercise on cognitive flexibility and hippocampus in aging male and female mice. Cognitive flexibility was assessed using reinstatement following extinction of conditioned fear. Mice were first conditioned with a tone conditioned stimulus (CS) that was paired with a footshock unconditioned stimulus (US), which led to freezing to the CS as a measure of emotional memory of the conditioning session (CS-US). Then the CS was presented repeatedly without the US, which decreases the freezing to the CS to form the extinction memory (CS- no US). When tested in the same context as extinction, the CS-no US memory is typically retrieved, evidenced by low levels of freezing. However, a single reminder footshock can facilitate the retrieval of the conditioning memory and lead to high freezing (i.e., reinstatement). Taken together, reinstatement can test cognitive flexibility because it requires flexible retrieval of the extinction vs conditioning memory depending on the reminder given (Anacker & Hen, 2017; Kim & Richardson, 2010; Short et al., 2016). While it 4 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.145136; this version posted June 11, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. is not the typical model to study cognitive flexibility, it is widely agreed that expression of reinstatement requires a complex understanding of environmental cues to be flexible in the choices of responses, and deficit in such cognitive flexibility may be related to persistence of fear observed in anxiety disorders (Bouton, 2002; Ganella & Kim, 2014; Giovanello, Schnyer, & Verfaellie, 2009; Kim & Richardson, 2010; Maren, Phan, & Liberzon, 2013). Consistent with these ideas, hippocampal lesions impair reinstatement (Frohardt, Guarraci, & Bouton, 2000; Wilson, Brooks, & Bouton, 1995). In those studies, it was not the context- specificity of reinstatement that was abolished, but reinstatement itself was abolished in the same context as where the reminder is given (Frohardt, Guarraci,